10 slides generated from the cancer page, with a quiz from the open benchmark and speaker notes that cite the sources. Arrow keys move between slides; Print gives one slide per page.
Unfavourable cancer of unknown primary is the large majority of cases, where a metastatic adenocarcinoma or poorly differentiated carcinoma fits no recognised pattern and its origin cannot be found. Treatment has long been general-purpose platinum chemotherapy, but CUPISCO showed that matching drugs to the tumour's genetic faults after short chemotherapy holds the disease longer.
Most patients with cancer of unknown primary do not fit a favourable subset. They have adenocarcinoma or poorly differentiated carcinoma in the liver, lungs, bones or several sites at once, are often unwell at diagnosis, and survival is short; performance status and serum lactate dehydrogenase are the strongest predictors of survival. For thirty years treatment has been empirical: carboplatin with paclitaxel or gemcitabine with cisplatin, regimens chosen because they work across many cancers, with response rates around a third and median survival of about nine to twelve months in trial populations and shorter in the clinic. Two randomised trials of classifier-directed site-specific chemotherapy (GEFCAPI 04, reported in 2019 as a conference abstract in Annals of Oncology, and a Japanese trial) did not improve on this, though the single-centre Fudan CUP-001 trial in China (Lancet Oncology 2024) found longer progression-free survival with site-specific therapy (a median of 9.6 against 6.6 months); the PD-1 antibody nivolumab produced responses in about a fifth of previously treated patients in the NivoCUP trial (Annals of Oncology 2022), leading to its approval for CUP in Japan in December 2021.
The CUPISCO trial (Lancet 2024) changed the framing. Of 636 patients enrolled with unfavourable non-squamous CUP, the 436 whose disease was controlled after three cycles of platinum-based induction chemotherapy were randomised three to one to switch to molecularly guided therapy (targeted drugs or immunotherapy chosen from comprehensive genomic profiling) or to continue chemotherapy: in this phase 2 trial median progression-free survival was 6.1 against 4.4 months (hazard ratio 0.72, 95 percent confidence interval 0.56 to 0.92), a modest but statistically significant gain that led its authors to recommend genomic profiling at diagnosis. About a third of patients carry an actionable alteration (HER2, BRAF V600E, NTRK fusions, MSI or high tumour mutational burden, and others), and the ESMO 2023 guideline recommends profiling for all patients fit for treatment; circulating tumour DNA is an alternative when tissue is scarce. Trials now test targeted agents, immunotherapy combinations and antibody-drug conjugates with chemotherapy in first line, and DNA-methylation classifiers are being revisited as a route to site-specific immunotherapy choices. Early palliative care and honest discussion of prognosis are part of standard management.
| Setting | Approach | Guideline |
|---|---|---|
| Work-up | Exclude favourable subsets; comprehensive genomic profiling of tissue or plasma; assess performance status and lactate dehydrogenase. | not mapped |
| First line, fit patients | Platinum-based doublet (carboplatin-paclitaxel or gemcitabine-cisplatin) for three cycles, then continuation or a switch to molecularly guided therapy where an actionable alteration is found (CUPISCO). | not mapped |
| Actionable alterations | Tumour-agnostic therapies: pembrolizumab or nivolumab for MSI-high or high tumour mutational burden, NTRK inhibitors, BRAF and MEK inhibitors, HER2-directed therapy. | not mapped |
| Second line | Alternative chemotherapy, PD-1 antibody if not given (nivolumab approved in Japan), or trial entry; trials of antibody-drug conjugates and immunotherapy combinations. | not mapped |
| Poor performance status | Best supportive care with early palliative care involvement. | not mapped |